Complex-Valued Spatio-Temporal Graph Convolution Neural Network optimized With Giraffe Kicking Optimization Algorithm...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Endocrinology

Background:

• Thyroid disorders significantly impact metabolism, posing diagnostic challenges.
• Accurate thyroid nodule classification is crucial for effective patient management.
• Existing diagnostic methods require enhancement for improved precision.

Purpose of the Study:

• To develop an advanced AI model for accurate thyroid nodule classification in ultrasound images.
• To improve the early detection of malignant thyroid nodules.
• To reduce misclassification rates and aid clinical decision-making.

Main Methods:

• Utilized a Complex-valued Spatio-Temporal Graph Convolution Neural Network (CSGCNN) optimized with Giraffe Kicking Optimization Algorithm (GKOA).
• Employed Bilinear Double-Order Filter (BDOF) for image pre-processing and Deep Adaptive Fuzzy Clustering (DAFC) for Region of Interest (RoI) segmentation.
• Extracted Geometric and Morphological features using Multi-Objective Matched Synchro Squeezing Chirplet Transform (MMSSCT).

Main Results:

• The proposed CSGCNN-GKOA-TNC-UI model demonstrated superior performance compared to existing methods.
• Achieved significant improvements in f-score and accuracy for thyroid nodule classification.
• The model showed enhanced accuracy in distinguishing between benign and malignant nodules.

Conclusions:

• The CSGCNN-GKOA-TNC-UI model offers a promising advancement in thyroid nodule classification.
• The approach aids radiologists and endocrinologists by enhancing diagnostic accuracy.
• Early malignancy detection and reduced unnecessary biopsies are key benefits of this AI model.